[Patch, fortran] PR34975 - [4.3 Regression] Bogus error with USEing modules

[Tobias Burnus]

Paul Richard Thomas wrote:
> Bootstrapped and regtested on x86_ia64/FC8 - OK for trunk?
>   
OK. Thanks for the fix, which fixes the compilation of ESPRESSO 
(www.quantum-espresso.org/), from which the test was "reduced" 
[fortunately the program is under the GPL which makes the inclusion in 
the test suite of this lengthy excerpt easier].


One note to symbol.c:

*************** gfc_undo_symbols (void)
*** 2582,2588 ****
        if (p->new)
  	{
  	  /* Symbol was new.  */
! 	  delete_symtree (&p->ns->sym_root, p->name);
  
  	  p->refs--;
  	  if (p->refs < 0)
--- 2582,2588 ----
        if (p->new)
  	{
  	  /* Symbol was new.  */
! 	  gfc_delete_symtree (&p->ns->sym_root, p->name);
  
  	  p->refs--;


I think you are using an out-dated GCC source. I have the following diff

@@ -2609,7 +2609,7 @@ gfc_undo_symbols (void)
                }
            }

-         delete_symtree (&p->ns->sym_root, p->name);
+         gfc_delete_symtree (&p->ns->sym_root, p->name);

          p->refs--;
          if (p->refs < 0)

  	  if (p->refs < 0)




Tobias